Underlying factors of social capital acquisition in the context of online-gaming: Comparing World of Warcraft and Counter-Strike

Computers in Human Behavior 36 (2014) 179–189

Contents lists available at ScienceDirect

Computers in Human Behavior journal homepage: www.elsevier.com/locate/comphumbeh

Underlying factors of social capital acquisition in the context of online-gaming: Comparing World of Warcraft and Counter-Strike Felix Reer a,⇑, Nicole C. Krämer b a b

University of Tuebingen, Department of Media Studies, Wilhelmstr. 50, 72074 Tuebingen, Germany University of Duisburg-Essen, Social Psychology: Media and Communication, Forsthausweg 2, 47057 Duisburg, Germany

a r t i c l e

i n f o

Article history:

Keywords: Social capital Self-disclosure Online-games Clans Guilds

a b s t r a c t The present study examines how players’ behaviors within gaming-communities (clans and guilds) influence the acquisition of social capital in online-gaming. In contrast to most existing studies, our study asks for crucial underlying factors of social capital acquisition and thereby includes players of online-games of different genres to allow comparisons. We hypothesize that frequently playing together (familiarity), participating in offline events (physical proximity) and being involved in clan/guild administration (social proximity) lead to more communication with fellow players and foster self-disclosure towards fellow players, which together facilitates the formation of bridging and bonding social capital. A sample of 682 clan/guild players of the games Counter-Strike and World of Warcraft was recruited via Internet and was asked to fill out a questionnaire. Results of a path analysis support our assumptions and enhance previous findings that players of online-games especially gain positive social outcomes, when they go beyond the game and join game-related groups, engage in clan/guild administration and participate in offline events. By revealing the crucial role of self-disclosure and communication frequency as underlying factors of social capital acquisition in online-gaming, our results provide a deeper insight into these mechanisms than existing studies. Our findings have implications of general importance, since the tested model worked well for player samples stemming from online-games of different genres. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction In recent years, computer- and videogames have more and more turned into a mass phenomenon. Simultaneously, they became a very important research area within media and communication studies and related fields like media psychology or media sociology. While early empirical research often concentrated on the negative outcomes of gaming like delinquency or aggressiveness (e.g. Anderson & Bushman, 2001; Dill & Dill, 1998; Gentile, Lynch, Linder, & Walsh, 2004), the upcoming of complex onlinegames like World of Warcraft (WOW) or Everquest changed the focus of research during the last years. In comparison to traditional videogames, online-games are social environments, where up to several thousands of players interact, communicate and play with each other. Therefore, more and more research has recently been conducted on social aspects and outcomes of playing. For example, social interactions with fellow players have often been identified as important motivation factors for playing online-games (e.g. Williams, Yee, & Caplan, 2008; Yee, 2006a, 2006b). Other studies ⇑ Corresponding author. Tel.: +49 7071 29 72351. E-mail address: [email protected] (F. Reer). http://dx.doi.org/10.1016/j.chb.2014.03.057 0747-5632/Ó 2014 Elsevier Ltd. All rights reserved.

addressed the potential of online-games to serve as starting points for network building and friendship formation (e.g. Cole & Griffiths, 2007; Götzenbrucker & Köhl, 2009; Shen, 2010). In the tradition of socio-scientific research on impacts of Internet usage (e.g. Kraut et al., 1998, 2002), our project focuses on social effects of using online-games. We are especially interested in the formation of social capital (Putnam, 2000; Williams, 2006b, 2007) in online-gaming and how it gets influenced by players’ behaviors within gaming-communities like clans and guilds. Although quite a few studies have already been conducted on social capital acquisition in gaming-worlds (e.g. Huvila, Holmberg, Ek, & Widen-Wulff, 2010; Steinkuehler & Williams, 2006; Williams, 2006a; Williams et al., 2006), only little is known about underlying psychological factors. Further, nearly all existing studies concentrate on only one specific game or games belonging to one specific genre, no attempts have been made to create a general model of social capital acquisition. In order to fill this gap, we constructed a path model with 5 underlying factors of social capital acquisition that is based on the work of Trepte, Reinecke, and Juechems (2012) and tested it with player samples stemming from two different games (World of Warcraft and Counter-Strike).

180

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

1.1. Social capital theory and the internet The concept of social capital has gained wide-spread attention in the past years and has frequently been referred to in scientific research of various fields (Trepte et al., 2012). It is often associated with Putnam (1995, 2000) who discussed disintegrative tendencies in the American society. Social capital can be defined as the benefits (for example information or emotional support) people obtain from their social networks (Ellison, Steinfield, & Lampe, 2007; Trepte et al., 2012; Williams, 2006b). In empirical research, authors often refer to Putnam’s distinction between two major types of social capital: Bridging social capital and bonding social capital (e.g. Ellison et al., 2007; Trepte et al., 2012; Williams, 2006b, 2007). Bridging social capital arises from rather tentative or casual relationships and may ‘‘[. . .] broaden social horizons or world views, or open up opportunities for information or new resources’’, but is commonly not useful in terms of gaining emotional support (Williams, 2006b, para. 14). Bonding social capital arises from deep or strong social ties that give emotional support and persist over a longer period of time, like close friends or family (Ellison et al., 2007; Trepte et al., 2012; Williams, 2006b). In his book ‘‘Bowling Alone’’ Putnam (2000) deals with the question, whether the increasing use of electronic media like television contributes to less time spent in social and civic activities and thereby threatens social capital. In the early years of the Internet, a lot of researchers asked similar questions about the social effects and outcomes of Internet usage. The findings of these studies were very diverse (e.g. Lee, 2009; Steinkuehler & Williams, 2006) and revealed negative (e.g. Kraut et al., 1998; Nie & Erbring, 2002; Nie, Hillygus, & Erbring, 2002), as well as positive (e.g. Kraut et al., 2002; Lee & Kuo, 2002) social effects of Internet usage. These inconsistencies in the findings can partly be explained by the fact that using the Internet allows a wide spectrum of activities, which makes asking for overall effects rather unsuitable (Steinkuehler & Williams, 2006). Accordingly, some forms of using the internet are fairly social and might have the potential to amplify social capital (e.g. writing e-mails, chatting), while others might displace social interactions and contribute to isolation (e.g. online shopping, online banking, viewing videos). As also pointed out by Trepte et al. (2012), recent studies take the diversity of the Internet into account and concentrate on effects of using specific services like social network sites (e.g. Burke, Kraut, & Marlow, 2011; Burke, Marlow, & Lento, 2010; Ellison et al., 2007; Sciandra, 2011; Valenzuela, Park, & Kee, 2009), chats and instant messengers (e.g. Bryant, Sanders-Jackson, & Smallwood, 2006; Lin, 2011), blogs (e.g. Ko & Kuo, 2009; Marlow, 2006) or onlinegames (e.g. Ducheneaut, Moore, & Nickell, 2007; Huvila et al., 2010; Steinkuehler & Williams, 2006; Williams, 2006a). For online-games, several studies demonstrated positive effects on bridging social capital (Trepte et al., 2012). For example the studies on massively multiplayer online-games (MMOs) conducted by Steinkuehler and Williams (2006) revealed that online-games are well suited for getting to know people with diverse worldviews and thereby especially foster bridging social capital. Williams et al. (2006) interviewed a representative sample of World of Warcraft players about their memberships in gaming-communities (guilds) and about the social outcomes they received. Players were found to use the game to meet new people and extend their social networks. Based on a survey study with Japanese participants Kobayashi (2010) argues that online-games could serve as sources for bridging social capital and foster social tolerance, since they have the potential to connect heterogeneous populations. The potential of online-gaming to create strong ties and bonding social capital is being discussed controversially (Trepte et al., 2012). Although most studies confirm the possibility of gathering bonding social capital in online-gaming, it has to be seen rather as an exception than

the normal case. For example Steinkuehler and Williams (2006) found some players of MMOs building up strong relationships; however these cases were very rare. Similar findings were reported by Williams et al. (2006) for World of Warcraft. Siitonen (2007) found that at least some players of online-gaming communities built up meaningful, strong relationships. Skoric and Kwan (2011) found a significant connection between the usage of MMOs and online bonding social capital, though the authors cannot completely rule out that this effect might be influenced by the specific setting of their study in the city-state of Singapore. However, none of the previously mentioned studies aimed at finding underlying psychological factors of social capital acquisition, all of them followed a more or less explorative attempt. In the following paragraph, we will concentrate on a study by Trepte et al. (2012), being the first effort to link social psychological knowledge and theories about friendship formation to social capital acquisition in the context of online-gaming. 1.2. Underlying factors of social capital acquisition in online-gaming Trepte et al. (2012) surveyed 811 members of online-gaming clans and developed a model of social capital acquisition in gaming-communities. Referring to social psychological research, they detected three underlying factors of social capital formation: physical proximity, social proximity and familiarity. Referring to Furnham (1989), Trepte et al. (2012) describe physical proximity as the ‘‘availability and accessibility of others for interaction’’ (p. 833). Physical closeness fosters individuals’ social affiliation and enhances the chances of friendship formation (Trepte et al., 2012; based on Fehr, 2008; Regan, 2011). Although ‘‘physical proximity does not guarantee attraction’’, ‘‘[. . .] affiliation and attraction are more likely to occur for physically proximate than for physically distant interaction partners.’’ (Trepte et al., 2012, p. 833) Trepte et al. (2012) argue that players of online-games are reachable for fellow players within the virtual world, ‘‘[. . .] but they usually are physically distant’’ (p. 833). For Steinkuehler and Williams (2006) this is a key argument, why playing online-games is less suitable in terms of gathering bonding social capital (Trepte et al., 2012). However, one should notice that several studies suggest the importance of offline events like LAN-parties as social facets of gaming, where players meet face-to-face for competing and playing together, but also for socializing and taking part in a wide spectrum of activities that go far beyond the mere gaming experience (e.g. Jansz & Martens, 2005; Jonsson & Verhagen, 2011; Taylor & Witkowski, 2010). Trepte et al. (2012) measured physical proximity by items referring to the participation in clan offline events. They found a positive relation between physical proximity and the development of bonding social capital, while bridging social capital was negatively associated with physical proximity. Trepte et al. (2012) suppose that online interactions more likely contribute to building up loose social ties, while participation in offline events might be especially important for building up deeper social bonds. Based on the work by Furnham (1989) and Regan (2011) Trepte et al. (2012) describe social proximity as ‘‘the closeness of social networks’’ (p. 834). Referring to social psychological literature, it can be argued that people who are easily accessible for many members of their social network (especially people in central positions within their networks), are more socially proximate, which fosters social attraction and affiliation (Trepte et al., 2012; based on Parks, 2007; Arriaga, Agnew, Capezza, & Lehmiller, 2008; Regan, 2011). Trepte et al. (2012) conclude that players who hold a central position within their online-gaming community and for example help in terms of administrating the group, generate more bridging and bonding social capital within their clans than less dedicated members. And indeed they found significant positive

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

connections between administrative involvement in a clan and the acquisition of both forms of social capital. Trepte et al. (2012) describe familiarity referring to ‘‘the social psychological notion that humans have a liking for the familiar’’ (p. 834; based on Furnham, 1989; Regan, 2011). Among other arguments, familiarity can be considered a factor of social capital acquisition since it ‘‘[. . .] promotes the acquisition of relationships, because of the interaction partners’ increasing similarity’’ (Trepte et al., 2012, p. 834, based on Regan, 2011). Trepte et al. (2012) assume that in the context of online-gaming, ‘‘[. . .] mechanisms that increase the players’ mutual encounters and contact frequency should be a good indicator of familiarity’’ (p. 836). Therefore they measured familiarity by the frequency players practice together with their clan mates. Practicing together (i.e. the training frequency) was found to be positively connected to clan bridging and bonding social capital. 1.2.1. Desiderata Although it can be evaluated as the most sophisticated approach so far, the study by Trepte et al. (2012) is certainly also characterized by several limitations. First, only members of ESLclans were included into the sample. The ESL (Electronic Sports League) is an organization that provides a platform where players can stick together easily and play clan matches. Similar to a sports league, the participating clans get ranked according to their performance. The ESL is especially popular among professional or semiprofessional players that are usually more ambitious and more involved than common players. Thus Trepte’s et al.’s (2012) results might only be valid for this special group of players. Second, the ESL is dominated by clans playing ego-shooter games like Counter-Strike or Call of Duty. But online-gaming communities are not a phenomenon limited to one specific genre. For example recent research especially pointed out the social aspects of guilds formed in Massively Online Roleplaying Games (MMORPGs) like World of Warcraft or Everquest (e.g. Williams et al., 2006). The model by Trepte et al. (2012) should therefore be tested with other types of games, making it possible to compare results. Third, Trepte et al. (2012) only identified three underlying factors of social capital acquisition. To get a more detailed insight into the phenomenon, more factors should be integrated into the model. It would especially be necessary to find out more precisely why and under which circumstances training frequency, meeting offline and involvement in clan administration lead to the formation of social capital. 2. Questions and hypotheses The first goal of our study is to enhance Trepte’s et al.’s (2012) model by adding two more factors that should be important for social capital acquisition in online-gaming: communication frequency and self-disclosure. Communicating with others is in general a crucial factor of getting into contact and building up relationships. In the context of Internet usage, the connections between communication frequency and friendship formation and/or maintenance have especially been examined regarding social network sites like Facebook. Gilbert and Karahalios (2009) presented a complex model on interpersonal networks within Facebook and found several communication-related measurements (e.g. exchange of wall posts and messages) predicted strength of social ties. In line with this, a study by Ledbetter et al. (2011) revealed communication frequency via Facebook as a predictor of relationship closeness, which confirmed prior findings from a study on the music-based social network site Last.fm (Baym & Ledbetter, 2009). Burke et al.

181

(2010) found that frequent communication with others via Facebook is connected with increased bonding social capital and decreased loneliness. In a more recent study, Burke et al. (2011) additionally found evidence that the frequency of receiving messages from Facebook friends is positively connected to bridging social capital formation. Self-disclosure is a term often used in social science and psychology, describing the degree to which people share personal information about themselves with others and to which degree they are willing to show their true self and for example talk about intimate and personal issues. Self-disclosure has often been identified as an important factor of relationship formation: For example Altman and Taylor (1973) argue that the development of strong and intimate relationships is based on growing levels of mutual self-disclosure. Thus, several empirical studies revealed positive connections between self-disclosure and liking one another and between self-disclosure and trust (e.g. Cozby, 1973; Wheeless & Grotz, 1977). Similar connections were found in studies on relationship formation and social capital acquisition in online-contexts: A study on Facebook by Sheldon (2009) found positive connections between self-disclosure, social attraction and trust. A path analysis as part of a study on privacy attitudes and behaviors of Facebook users showed positive connections between users’ willingness to disclose on Facebook and the perception of bridging and bonding social capital (Stutzman, Vitak, Ellison, Gray, & Lampe, 2012). Ko and Kuo (2009) found connections between bloggers’ amount of self-disclosure and the bridging and bonding social capital they received. For our study we expect that meeting clan/guild mates offline (physical proximity), frequently playing together (familiarity) and being involved in clan/guild administration (social proximity) lead to higher levels of communication frequency and self-disclosure and thereby foster the formation of bridging and bonding social capital. These assumptions are taken together in a path model (see Fig. 1). In the following we will explain the different paths and the corresponding hypotheses. 2.1. Physical proximity Participating in social events of a clan/guild (physical proximity) and getting the chance to interact with mates in a face-to-face situation should automatically facilitate communication. Our first hypothesis therefore reads as follows: H1a. Participation in offline events of a clan/guild is positively related to communication frequency. Meeting clan/guild-mates physically also automatically implies revealing more information about one’s own self. Besides the possibility to talk face-to-face, stepping from the anonymous onlinecontext to the more personal offline world always includes the possibility to gather additional information about the interactionpartner and for example see how he/she looks, behaves and reacts. Thus meeting offline should foster trust and intimacy and thereby also increase the willingness of verbal self-disclosure. Hence we hypothesize: H1b. Participation in offline events of a clan/guild is positively related to self-disclosure. While building up loose relationships seems to work well in the online-context, meeting in the real world already proved to be a crucial factor for the formation of meaningful, deep relationships. As mentioned, the positive effects of meeting clan/guild members offline should not be limited to an increase in verbal self-disclosure, but also foster trust and intimacy by reducing anonymity.

182

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

Fig. 1. Predicted path model with hypotheses.

We therefore think that meeting other clan/guild mates physically should have an additional direct impact on the formation of deep relationships (bonding social capital): H1c. Participation in offline events of a clan/guild is positively related to bonding social capital. 2.2. Familiarity Nowadays most online-games provide their players with the possibility to directly communicate with each other using chats or voice functions while playing. Additionally, many clans/guilds use voice over IP solutions like Teamspeak or Skype when playing. Thus it can be argued, that frequently playing together with clan/ guild mates (familiarity) is accompanied by an increase of communication: H2a. Frequency of playing together with clan/guild mates is positively related to communication frequency. Sticking together in a team regularly and trying to reach the game’s goals together should enhance solidarity and trust towards fellow players. Frequency of playing together with clan/guild mates is therefore expected to foster players’ willingness to selfdisclose: H2b. Frequency of playing together with clan/guild mates is positively related to self-disclosure. 2.3. Social proximity Online-gaming communities like clans and guilds typically organize themselves similar to sports clubs: a group of players manages the team and cares for administrative tasks like arranging trainings or offline-meetings. In many cases the leader of the clan or the master of the guild is responsible for recruiting new members, organizing matches with other groups and solving conflicts within the group. A web-master or server-master is responsible

for technical aspects like the administration of the clan/guild homepage or game-servers. As already pointed out, members involved in such administrative tasks are expected to have a central position within the group and to serve as contact persons for the common members. Hence involvement in the administration or management of a clan/guild (social proximity) should be related to an increase in communication with others: H3a. Involvement in clan/guild administration is positively related to communication frequency. Clan-leaders, guild-masters or server-masters are crucial for the functioning and cohesion of online-gaming communities. They take responsibility for the continued existence of the group. Normally these positions are taken by experienced players who are members of the group for several years and highly identify with their clan/guild. They are typically the most well-known and trusted persons within their clan/guild. Thus we suppose, getting accepted and trusted as one of the leaders or administrators of the group requires offering more information about one’s own person than common players normally do: H3b. Involvement in clan/guild administration is positively related to self-disclosure. The effects of taking over a leading position in a clan/guild should not be limited to more communication and more self-disclosure. Becoming one of the clan/guild managers also enhances prestige and reputation within the group and thereby makes one more attractive to other group members. Hence we think the involvement in clan/guild administration should also directly influence the chances of building up relationships to fellow players: H3c. Involvement in clan/guild administration is positively related to bonding social capital.

H3d. Involvement in clan/guild administration is positively related to bridging social capital.

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

2.4. Communication/self-disclosure As mentioned above, several studies found communication frequency and self-disclosure being positively related to the formation of bridging and bonding social capital (e.g. Burke et al., 2010, 2011; Ko & Kuo, 2009). Hence we assume: H4a. Communication frequency is positively related to bonding social capital. H4b. Communication frequency is positively related to bridging social capital.

183

killer-games, the social aspects of CS remained understudied so far. World of Warcraft is one of the most popular MMORPGs of the last years with more than 12 million active players. Both games are typical multiplayer-games, characterized by facilitating cooperative play and thereby fostering the formation of persistent online-gaming communities named clans (for CS) and guilds (for WOW). To reach a wide spectrum of players, we spread hyperlinks to the survey via several channels including game-related forums, Facebook interest groups and online-communities. To make participating more attractive, we raffled three 50 Euro coupons for a popular online store among all participants. All questions were asked in German and the survey was accessible for a period of 6 weeks. 3.1. Sample

H5a. Self-disclosure is positively related to bonding social capital. H5b. Self-disclosure is positively related to bridging social capital. 2.5. Additional effects/questions We additionally suppose that there are significant indirect effects of meeting offline (physical proximity), playing together (familiarity) and being involved in clan/guild administration (social proximity) via communication frequency and self-disclosure on bridging and bonding social capital. To test these indirect effects, we hypothesize: H6a. Participation in offline events of a clan/guild indirectly influences the formation of bridging and bonding social capital via communication frequency and self-disclosure. H6b. Frequency of playing together with clan/guild mates indirectly influences the formation of bridging and bonding social capital via communication frequency and self-disclosure. H6c. Involvement in clan/guild administration indirectly influences the formation of bridging and bonding social capital via communication frequency and self-disclosure. One of the desiderata of Trepte’s et al.’s (2012) study relates to the restriction on ESL-players. Since the ESL is dominated by egoshooter games, another main goal of our study is to test the model with players of games belonging to different genres. Since our model is based on social psychological findings and theories which should be meaningful for any type of online-game, we assume that the model should work well and generate similar results for games belonging to different game genres. We therefore hypothesize that there are no significant differences in the strengths of the path weights between player samples stemming from different types of games:

All in all, 682 clan/guild-players filled out our questionnaire (391 WOW players, 291 CS players). As typical for games research, a large majority (90.2%) of the participants were male (N = 615). Age ranged from 14 to 73 years with a mean of 23.91 years (SD = 7.68). On average, participants spend 20.40 hours per week with playing (SD = 14.48) and were members of their clan/guild for 31.08 months (SD = 28.73). 3.2. Measurements 3.2.1. Bridging and bonding social capital To measure social capital 18 items of Williams’ (2006b) online social capital scale were translated into German and edited to better fit the gaming context (e.g. ‘‘Interacting with fellow players makes me interested in things that happen outside of my town.’’). Participants were asked to rate the items on a five-point scale ranging from 1 ‘‘strongly disagree’’ to 5 ‘‘strongly agree’’ and thereby think of persons they got to know by playing CS (for the CS sample) or WOW (for the WOW sample). 9 items focus on deep relationships (bonding social capital) while the other 9 items relate to weak relationships (bridging social capital). Cronbach’s a for both subscales was acceptable with a = .848 for bridging (M = 3.38; SD = .80) and a = .865 for bonding (M = 3.02; SD = .95) social capital. We summed up the scores for both subscales and averaged the results for further steps. 3.2.2. Offline activities (physical proximity) Two items of the questionnaire focus on players’ engagement in offline clan activities (‘‘The members of my clan/guild also regularly meet outside the internet’’; ‘‘I have met quite a lot of the members of my clan/guild in real life’’). A five-point scale was used ranging from 1 ‘‘strongly disagree’’ to 5 ‘‘strongly agree’’. The summed up and averaged scores were saved for further analysis (M = 2.97; SD = 1.30; a = .764).

3. Material and methods

3.2.3. Administrative involvement (social proximity) Another two items with five-point scales ranging from 1 ‘‘strongly disagree’’ to 5 ‘‘strongly agree’’ relate to players’ involvement in the administration of their clan or guild (‘‘In my clan/guild I spend a lot of time on administrative or technical things’’; ‘‘I’m not only playing, but also take care of administrative issues of my clan/guild’’). Mean for the summed up and averaged score is M = 3.11 (SD = 1.23; a = .807).

We implemented a questionnaire into an online-survey system and made it accessible in the Internet. To allow comparisons between different game-genres, we recruited players of two different popular online-games: Counter-Strike (CS) and World of Warcraft (WOW). Counter-Strike belongs to the genre of ego-shooters and has often been discussed in the context of school-shootings. Possibly because of its bad image as one of the most played

3.2.4. Playing together (familiarity) Trepte et al. (2012) measured familiarity by asking clan-players about their frequency of training together. Training for matches is typical for professional or semi-professional ESL-Clans. For fun orientated groups training should not be a good indicator of familiarity. Also, training is typical for shooter-games or strategy games but does not play such an important role within MMORPGs. There-

H7. The path weights of the model do not significantly differ in their strengths between player samples stemming from onlinegames of different genres.

184

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

fore we decided to measure clan/guild mates’ frequency of playing together instead of their training frequency. Participants were asked to estimate the number of days per month they meet online with clan/guild mates to play together (‘‘To play together we meet online on __ days per month’’). The answers range from 0 to 30 with an average of M = 14.20 (SD = 8.46).

4.1. Results for World of Warcraft As shown in Table 1, we found significant correlations between offline activities, playing together and administrative involvement. We therefore allowed all three variables to co-vary in the model. For the same reason we allowed the error terms of communication frequency and self-disclosure as well as the error terms of bridging and bonding social capital to co-vary. The calculated model for WOW is shown in Fig. 2. The model explains 26% of the variance in self-disclosure (R2 = .26), 24% of the variance in communication frequency (R2 = .24), 55% of the variance in bonding social capital (R2 = .55) and 37% of the variance in bridging social capital (R2 = .37).

3.2.5. Communication frequency The frequency of communication with fellow players was measured with twelve items concerning the usage of different communication channels (‘‘How often do you communicate with fellow players using the following communication channels?’’). We used a five-point scale ranging from 1 ‘‘never’’ to 5 ‘‘very often’’ and took text-based (e.g. in game chat, instant messenger), voice-based (e.g. Teamspeak, Skype) as well as non-computer mediated (e.g. telephone, face-to-face) ways of communicating into consideration. Scores were summed up and averaged to build an indicator of overall communication frequency (M = 2.85; SD = .67; a = .748).

4.1.1. Model fit Like Trepte et al. (2012), we computed several established fit criteria to evaluate model fit: The model was accepted with a root mean square error of approximation (RMSEA) below .06 (Hu & Bentler, 1999), a comparative fit index (CFI) above .95 (Hu & Bentler, 1999) and a relative chi-square value (CMIN/df) below 2.00 (Byrne, 1989). The model fits the data excellently with RMSEA = .00 (90% confidence interval from .000 to .067), CFI = 1.00 and CMIN/df = .528.

3.2.6. Self-disclosure Existing scales regarding self-disclosure are rather complex and include several sub-dimensions of the construct (e.g. Leung, 2002; Wheeless, 1976; Wheeless & Grotz, 1976). Since these scales were too long for our survey and also did not fit the gaming context very well, we decided not to adopt them completely, but to create 7 new items which were partly orientated on items used by Wheeless (1976) and Leung (2002). These statements include the willingness to disclose the real self (‘‘When talking with fellow players I openly show who I really am.’’), to talk about intimate and private things (‘‘I sometimes talk with fellow players about very personal and intimate things.’’) to reveal personal opinions and believes (‘‘Only infrequently do I express my personal opinions and believes towards fellow players’’ reverse) and to reveal negative aspects of the own self (‘‘I also disclose undesirable things about myself towards fellow players’’). Additional items focus on talking about job issues or family issues (e.g. ‘‘I sometimes talk with fellow players about problems I have at work or school’’). All statements were provided with a five-point scale ranging from 1 ‘‘strongly disagree’’ to 5 ‘‘strongly agree’’. As the a-value for the scale was within acceptable range (a = .810), scores were summed up and averaged for further operations (M = 2.97; SD = .85).

4.1.2. Direct effects As predicted in H1a, we found a positive connection between offline activities and communication frequency (b = .33; p < .001). H1b predicted a positive relation between offline activities and self-disclosure and can be accepted with b = .14 (p < .01). Conforming H1c, a significant direct link between offline activities and bonding social capital (b = .09; p < .05) was found. H2a/b predicted positive relations between frequency of playing together and communication frequency and between frequency of playing together and self-disclosure. The data confirms both predictions with b = .09 (p < .05) for communication frequency and b = .17 (p < .001) for self-disclosure. Approving H3a/b/c/d, involvement in guild administration was positively related to communication frequency (b = .20; p < .001) and self-disclosure (b = .35; p < .001) and additionally was directly linked to bonding (b = .14; p < .001) and bridging (b = .18; p < .001) social capital. As predicted by H4a/b, communication frequency was positively related to both types of social capital (bonding: b = .13; p = .001; bridging: b = .25; p < .001). Also self-disclosure (H5a/b) was found to be positively related to bonding (b = .55; p < .001) and bridging (b = .34; p < .001) social capital.

4. Results As a first step, we calculated correlations, means and standard derivations for all variables and both player groups (CS and WOW). Results are shown in Tables 1 and 2. To analyze hypothesis H1–H5, we performed path analysis (maximum likelihood method) using AMOS 20 software package. To analyze the predicted indirect effects (H6a/b/c) we used bootstrapping (e.g. Bühner & Ziegler, 2009; Preacher & Hayes, 2008) with 2000 bootstrap samples and a bias-correlated confidence level of 90%.

4.1.3. Indirect effects H6a/b/c predicted indirect effects of offline activities, playing together and administrative involvement on bridging and bonding social capital, via self-disclosure and communication frequency. To test these hypotheses, we used the bootstrapping function (2000 bootstrap samples) of AMOS. All three predictions were affirmed: Meeting offline (bridging: b = .13; p < .001; 90% bias-corrected con-

Table 1 Means, standard deviations and Pearson-correlations for the WOW sample.

1. 2. 3. 4. 5. 6. 7. *

Offline-activities Playing together Admin. involvement Self-disclosure Comm. frequency Bridging social capital Bonding social capital

p < .05. ** p < .01.

M

SD

1

2

3

4

5

6

7

2.60 13.57 2.91 2.95 2.71 3.36 2.98

1.25 8.45 1.23 .91 .70 .83 .97

– .198** .409** .319** .432** .271** .374**

– .320** .306** .226** .258** .264**

– .459** .370** .425** .481**

– .413** .524** .702**

– .455** .448**

– .570**

–

185

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189 Table 2 Means, standard deviations and Pearson-correlations for the CS sample.

1. 2. 3. 4. 5. 6. 7. * **

Offline-activities Playing together Admin. involvement Self-disclosure Comm. frequency Bridging social capital Bonding social capital

M

SD

1

2

3

4

5

6

7

3.46 15.04 3.37 3.00 3.03 3.39 3.09

1.21 8.41 1.17 .76 .59 .76 .91

– .053 .070 .204** .238** .066 .354**

– .187** .145* .136* .139* .084

– .214** .261** .271** .277**

– .207** .309** .621**

– .264** .250**

– .415**

–

p < .05. p < .01.

Fig. 2. Path model for the WOW sample. All reported standardized beta coefficients are significant at least with p < .05.

fidence interval from .09 to .18; bonding: b = .12; p < .001; 90% bias-corrected confidence interval from .07 to .18), regularly playing together (bridging: b = .08; p < .001; 90% bias-corrected confidence interval from .05 to .12; bonding: b = .10; p < .001; 90% bias-corrected confidence interval from .06 to .15) and administrative involvement (bridging: b = .17; p < .001; 90% bias-corrected confidence interval from .13 to .22; bonding: b = .22; p < .001; 90% bias-corrected confidence interval from .17 to .28) indirectly influenced social capital acquisition via communication frequency and self-disclosure.

together were significantly related to each other and therefore were allowed to co-vary in the model. Additionally, we found significant correlations between communication frequency and selfdisclosure and between bridging and bonding social capital. Thus we allowed the corresponding error terms to co-vary as well. The resulting model for CS is presented in Fig. 3. The model explains 9% of the variance in self-disclosure (R2 = .09), 12% of the variance in communication frequency (R2 = .12), 46% of the variance in bonding social capital (R2 = .46) and 16% of the variance in bridging social capital (R2 = .16).

4.2. Results for Counter-Strike Table 2 shows the results of the correlation analysis for the CS sample. Administrative involvement and frequency of playing

4.2.1. Model fit As for the WOW sample, we calculated RMSEA, CFI and CMIN/df to evaluate model fit. The model proved to fit the data very well

186

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

Fig. 3. Path model for the CS sample. Untagged reported standardized beta coefficients are significant at least with p < .05. *p = .059. **p = .275.

with RMSEA = .00 (90% confidence interval from .000 to .076); CFI = 1.00 and CMIN/df = .865. 4.2.2. Direct effects Offline activities were positively related to communication frequency (b = .23; p < .001) and self-disclosure (b = .20; p < .001). Additionally, a significant direct link between offline activities and bonding social capital (b = .23; p < .001) was found. Thus H1a/b/c can be accepted. As predicted by H2a/b, frequency of playing together was significantly related to communication frequency (b = .11; p = .059) and self-disclosure (b = .12; p < .05). Hypothesis H3a/b/c/d can also be accepted: Administrative involvement was significantly related to communication frequency (b = .23; p < .001), self-disclosure (b = .18; p < .01), bonding (b = .13; p < .01) and bridging (b = .18; p < .01) social capital. Self-disclosure was a significant predictor of bonding (b = .54; p < .001) and bridging (b = .24; p < .001) social capital, therefore H5a/b were also confirmed. H4b predicted a positive relation between communication frequency and bridging social capital and can be accepted with b = .17 (p < .01). But there was no significant relation between communication frequency and bonding social capital (b = .05; p = .275), thus H4a must be rejected. 4.2.3. Indirect effects Regarding H6a/b/c we bootstrapped (2000 bootstrap samples) indirect effects of offline activities, playing together with clan mates and administrative involvement on bridging and bonding social capital via self-disclosure and communication frequency. Equally to the WOW sample, all predicted indirect links were significant. Bridging and bonding social capital get indirectly influenced by offline activities (bridging: b = .09; p < .001; 90% biascorrected confidence interval from .05 to .12; bonding: b = .12; p < .001; 90% bias-corrected confidence interval from .07 to .17),

frequency of playing together with clan mates (bridging: b = .05; p < .01; 90% bias-corrected confidence interval from .02 to .08; bonding: b = .07; p < .05; 90% bias-corrected confidence interval from .02 to .12) and administrative involvement (bridging: b = .08; p < .001; 90% bias-corrected confidence interval from .05 to .13; bonding: b = .11; p < .001; 90% bias-corrected confidence interval from .05 to .17).

4.3. Comparing World of Warcraft and Counter-Strike The path model works well for both games. Even though our model was based on a model originally designed for ESL-clans and mainly ego-shooter games, it explains more variance and fits the data slightly better for the WOW sample. H7 assumed that the strengths of the path weights do not significantly differ between the two player samples. We used the multi-group analysis function of Amos to compare the strengths of the path weights between the two samples. Since this procedure is done by comparing the same model across groups, the model for the CS sample had to be slightly adjusted beforehand: Whereas within the WOW sample, we found significant correlations between offline activities and playing together, as well as between offline activities and administrative involvement, these correlations were not present within the CS sample. To solve this problem and to equalize the models, we added the missing correlations to the CS model, but manually fixed them to zero (which is statistically the same thing as to leave them out). After this preparation, we ran a multi-group analysis and examined the significance of the differences in the strengths of the path weights between the groups by investigating the critical ratios (CR) for pairwise parameter comparisons (a value of CR > 1.645 indicates significance with p < .1, a value of CR > 1.96 indicates significance with p < .05 and a value of CR > 2.58 indicates significance with p < .01). The results are shown in Table 3.

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189 Table 3 Pairwise parameter comparisons of the path weights. Path

H1a H1b H1c H2a H2b H3a H3b H3c H3d H4a H4b H5a H5b

WOW sample

CS sample

CR

B

b

B

b

.19 .11 .07 .01 .02 .12 .26 .11 .12 .18 .30 .59 .31

.33 .14 .09 .09 .17 .20 .35 .14 .18 .13 .25 .55 .34

.11 .12 .17 .01 .01 .11 .11 .10 .11 .08 .22 .64 .24

.23 .20 .23 .11 .12 .23 .18 .13 .18 .05 .17 .54 .24

1.945 .370 2.407 .092 .975 .097 2.745 .207 .112 1.117 .855 .718 1.051

Sig.

*

n.s. **

n.s. n.s. n.s.

187

orientated than WOW guilds. Thus it might be assumed that communication in the context of CS is often less socially orientated and more task-orientated and therefore does not necessarily foster the formation of bonding social capital. Altogether H7 can partly be affirmed: Most path weights do not significantly differ between the two player samples and in general the model works well for both games. However, three path weights showed significant differences which we tried to explain by game specific characteristics.

***

n.s. n.s. n.s. n.s. n.s. n.s.

*

p < .1. p < .05. *** p < .01. **

As expected, most path weights do not significantly differ between the player samples. However, three path weights show significant differences in their strengths. In the WOW sample, participation in offline-activities is significantly stronger linked to communication frequency than in the CS sample (b = .33 for WOW and b = .23 for CS; CR = 1.945; p < .1). Instead there is a significantly stronger direct effect of meeting offline on bonding social capital within the CS sample (b = .23 for CS and b = .09 for WOW; CR = 2.407; p < .05). This might be due to the fact that WOW guilds normally consist of a higher number of players than CS clans. One might conclude that also clan and guild meetings differ in the number of participating players. Large events should offer more opportunities to talk with many people, but smaller events might be better to intensively get to know each other and thereby form bonding social capital. Administrative involvement is significantly more strongly linked to self-disclosure within the WOW sample (b = .35 for WOW and b = .18 for CS; CR = 2.745; p < .01). This might be explained with differences in the structures of WOW guilds and CS clans: WOW provides guilds with a basic technical infrastructure within the game and therefore makes it easier for guilds to organize themselves. Helping in the administration of a guild therefore should mainly mean leading and managing the group. CS clans need to care for the technical infrastructure themselves; there is no in-game support for launching clans. Thus within CS clans, a lot of people are involved in technical tasks like setting up game-servers, forums or homepages. These lower level tasks might not require such high levels of trust and faith like getting into the position of a WOW guild master and hence do not necessarily go hand in hand with more selfdisclosure. Even though the differences in the strengths of these path weights are not significant, it is also noticeable that communication frequency is not significantly linked to bonding social capital and less strongly linked to bridging social capital within the CS sample (bonding: b = .13 for WOW and b = .05 for CS; bridging: b = .25 for WOW and b = .17 for CS). This might be explained by the fact that the playing situation in CS is different to the playing situation in WOW. Playing WOW includes situations where you are less involved and spend time with rather undemanding tasks like leveling. These situations might leave more room for serious conversations with fellow players, while playing CS is very action-orientated and communication while playing is typically dominated by short and task-orientated conversations. Also CS clans often see themselves as modern forms of sport-clubs and especially professional clans might in general be less socially

5. Discussion and conclusions The aim of our study was to find underlying factors of social capital acquisition in the context of online-gaming and thereby enhance the findings of Trepte et al. (2012). While Trepte et al. (2012) tested the direct effects of physical proximity, social proximity and familiarity on bridging and bonding clan social capital and additionally tested the direct and indirect effects of these constructs on offline social support, we concentrated on the question, how and why physical proximity, social proximity and familiarity effect the acquisition of social capital in online gaming. Therefore we added self-disclosure and communication frequency as additional factors, resulting in a model of five underlying factors of social capital acquisition. Another main goal of our study was to test the model with player samples stemming from different types of games and thereby allow comparisons. While Trepte et al. (2012) recruited their player sample on the website of the ESL and therefore mainly reached players of ego-shooter games, we spread the links to our survey on several channels and recruited players of a popular shooter-game (Counter-Strike) as well as players of a popular MMORPG (World of Warcraft). In both player samples physical proximity (measured by participation in clan/guild offline meetings), social proximity (measured by involvement in clan/guild administration) and familiarity (measured by frequency of playing together with clan/guild mates) proved to be important predictors of communication frequency and self-disclosure. Thus it can be concluded that players participating in offline events of their clan/guild, frequently playing together with clan/guild mates and being involved in administrative tasks of their clan/guild communicate more with fellow players and also disclose more personal information towards fellow players. Communication frequency was significantly linked to bridging social capital in both samples and additionally to bonding social capital in the WOW sample. To communicate a lot with fellow players therefore proved to be crucial for gathering social capital within online-gaming. Self-disclosure had a rather strong effect on the formation of bonding social capital and a medium effect on the formation of bridging social capital in both samples. Hence it can be assumed that players who self-disclose a lot increase their chances of gathering social capital, with self-disclosure seeming to be particularly important for the formation of bonding social capital. Additionally, significant indirect effects of meeting offline, playing together and administrative involvement on bridging and bonding social capital via communication frequency and self-disclosure were found. Taken together our findings support Trepte’s et al.’s (2012) findings that playing online-games especially leads to positive social outcomes, when players do more than just playing and engage in game-related social groups. Furthermore, the revealed indirect effects amend Trepte’s et al.’s (2012) results in the sense that playing in a guild or clan, being involved in its management and participating in offline events fosters communication with fellow players, enhances the willingness for self-disclosure and thereby increases the chances of gathering social capital. With these findings our study provides a deeper insight into the underlying mech-

188

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189

anisms of social capital formation in online-gaming than previous studies and emphasizes the necessity of taking additional factors into account when asking for social outcomes of gaming. Looking for one-dimensional overall effects is not suitable, since outcomes differ depending on the way people use games, the way they act towards fellow players and the way they are integrated in gamerelated groups and structures. Our model proved to fit the data well for both games, even though these games belong to different game genres. A multigroup analysis with pairwise comparisons of the path weights revealed that only three of the thirteen path weights within the model significantly differed between the groups. Noteworthy, all compared path weights pointed in the same direction and the significant differences were only based on slight varieties in the strengths of the path weights. It therefore can be carefully assumed that the discovered factors are of general importance for the acquisition of social capital within online-gaming and should also be affirmed by data stemming from other games and genres. The results also confirm the adequacy of social psychological theories for research on media effects: the positive impacts of social proximity, physical proximity, familiarity, communication frequency and self-disclosure on friendship formation have been well-known in psychology for a long period of time and in our study also proved to be of importance in the context of virtual environments like online-gaming communities. In this, our findings underline Trepte’s et al.’s (2012) notion that social psychological offline concepts have the potential to provide valuable insights into friendship formation within online settings. It should be fruitful to follow this route in future research on outcomes of gaming. 5.1. Limitations and future steps First, like Trepte et al. (2012), we chose a cross-sectional design for our study. The directions of the detected paths are plausible against the background of social psychological theories, but should be confirmed by long-term investigations of gaming-effects. The second limitation concerns the data-samples used in this study. Participants were recruited by spreading links in the Internet. This self-recruitment might not be adequate to produce a representative sample which could possibly influence the results. Even though our sample-sizes were large enough to get an impression of the investigated effects, our results should be validated with larger and more representative samples. Our samples consisted solely of German players. Even though the results of some studies on Asian players of online-games (e.g. Kobayashi, 2010; Skoric & Kwan, 2011) give first evidence that social capital acquisition in online-gaming generally could take place in any type of online-gaming context (regardless of the cultural backgrounds of the players), there is a high need for systematical, cross-cultural comparisons regarding the underlying factors of social capital acquisition. Filling this gap and testing our model with player samples stemming from different countries and cultural backgrounds may be a promising goal for further studies. Additionally, although we demonstrated that our model works fine for the two investigated games, the model should be tested with more games of different genres in order to allow a serious general statement. For example many strategy games also foster the emergence of gaming-communities (called alliances) that should potentially evoke similar player behaviors and social effects. Adapting our model and the revealed underlying factors of social capital acquisition for new forms of online-games like browser-games or social media games might also be an interesting task for future research. Third, the measurements used in this study underlie some limitations. To keep the survey short, we used only two items to measure involvement in clan/guild management and two items to measure participation in offline events. Future studies should try

to more precisely distinguish between different forms of offline events and different forms of administrative involvement. For example participating in game-related mass events like huge LAN-parties might have other effects than meeting clan/guild mates at smaller and more intimate, socially-oriented events. It also might make a difference what kind of administrative position someone holds in his/her clan/guild. Taking a leading position might generate more positive outcomes than holding an inferior technical position. In our study we measured overall communication frequency by summing up the frequencies of using several forms of communication, including face-to-face communication, voice-based forms of communication and text-based forms of communication. Further analysis should take a closer look at the effects of each of these different forms of communication. It seems logical, that choosing more personal forms of communication like face-toface or voice-based communication is especially important for the development of bonding social capital, while the usage of textbased forms of communication might be linked more strongly to the emergence of bridging social capital. For economic reasons, we only used 7 items to measure players’ willingness to self-disclose towards fellow players. Even though we tried to include a wide spectrum of statements, our self-disclosure scale was less complex than other instruments used in this context and did not consist of multiple sub-dimensions. Several of our items focused on facets of revealing private, personal and intimate things towards fellow players. It makes sense that this aspect of self-disclosure is especially crucial for the emergence of strong social ties, which is supported by the fact that one item of the bonding social capital scale adopted from Williams (2006b) focusses on talking about intimate personal problems. Another interesting task for future studies might therefore be to use a more complex measurement of self-disclosure and to separately discover the influences of different aspects and sub-dimensions of the construct on the formation of both forms of social capital. References Altman, I., & Taylor, D. A. (1973). Social penetration: The development of interpersonal relationships. New York: Holt, Rinehart and Winston. Anderson, C. A., & Bushman, B. J. (2001). Effects of violent video games on aggressive behavior, aggressive cognition, aggressive affect, physiological arousal, and prosocial behavior: A meta-analytic review of the scientific literature. Psychological Science, 12(5), 353–359. Arriaga, X. B., Agnew, C. R., Capezza, N. M., & Lehmiller, J. J. (2008). The social and physical environment of relationship initiation: An interdependence analysis. In A. Wenzel, J. H. Harvey, & S. Sprecher (Eds.), Handbook of relationship initiation (pp. 27–41). New York: Psychology Press. Baym, N. K., & Ledbetter, A. (2009). Tunes that bind?: Predicting friendship strength in a music-based social network. Information, Communication & Society, 12(3), 408–427. Bryant, J. A., Sanders-Jackson, A., & Smallwood, A. M. (2006). IMing, text messaging, and adolescent social networks. Journal of Computer-Mediated Communication, 11(2), 577–592. Bühner, M., & Ziegler, M. (2009). Statistik für Psychologen und Sozialwissenschaftler [Statistics for psychologists and social scientists]. Munich, Germany: Pearson Studium. Burke, M., Marlow, C., & Lento, T. (2010). Social network activity and social wellbeing. In CHI ‘10 Proceedings of the SIGCHI conference on human factors in, computing systems (pp. 1909–1912). Burke, M., Kraut, R., & Marlow, C. (2011). Social capital on Facebook: Differentiating uses and users. In CHI ‘11 proceedings of the SIGCHI conference on human factors in, computing systems (pp. 571–580). Byrne, B. M. (1989). A primer of LISREL: Basic applications and programming for confirmatory factor analytic models. New York: Springer. Cole, H., & Griffiths, M. D. (2007). Social interactions in massively multiplayer online role-playing gamers. CyberPsychology & Behavior, 10(4), 575–583. Cozby, P. C. (1973). Self-disclosure. A literature review. Psychological Bulletin, 79(2), 73–91. Dill, K. E., & Dill, J. C. (1998). Video game violence: A review of the empirical literature. Aggression and Violent Behavior, 3(4), 407–428. Ducheneaut, N., Moore, R., & Nickell, E. (2007). Virtual third places: A case study of sociability in massively multiplayer games. Computer Supported Cooperative Work, 16(1–2), 129–166. Ellison, N., Steinfield, C., & Lampe, C. (2007). The benefits of facebook ‘‘friends:’’ Social capital and college students’ use of online social network sites. Journal of

F. Reer, N.C. Krämer / Computers in Human Behavior 36 (2014) 179–189 Computer-Mediated Communication, 12(4). . Fehr, B. (2008). Friendship formation. In A. Wenzel, J. H. Harvey, & S. Sprecher (Eds.), Handbook of relationship initiation (pp. 29–54). New York: Psychology Press. Furnham, A. (1989). Friendship and personal development. In R. Porter & S. Tomaselli (Eds.), The dialectics of friendship (pp. 92–110). London: Routledge. Gentile, D. A., Lynch, P. J., Linder, J. R., & Walsh, D. A. (2004). The effects of violent video game habits on adolescent aggressive attitudes and behaviors. Journal of Adolescence, 27(1), 5–22. Gilbert, E., & Karahalios, K. (2009). Predicting tie strength with social media. In CHI ‘09 proceedings of the SIGCHI conference on human factors in computing systems (pp. 211–220). Götzenbrucker, G., & Köhl, M. (2009). Ten years later. Towards the careers of longterm gamers in Austria. Eludamos. Journal for Computer Game Culture, 3(2), 309–324. Hu, L.-T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. Huvila, I., Holmberg, K., Ek, S., & Widen-Wulff, G. (2010). Social capital in second life. Online Information Review, 34(2), 295–316. Jansz, J., & Martens, L. (2005). Gaming at a LAN event: The social context of playing video games. New Media & Society, 7(3), 333–355. Jonsson, F., & Verhagen, H. (2011). Sensing game play. Exploring computer game play in a game café and a mass LAN party. In Proceeding of the 16th international conference on computer games, CGAMES 2011 (pp. 134–141). Ko, H. C., & Kuo, F. Y. (2009). Can blogging enhance subjective well-being through self-disclosure? CyberPsychology & Behavior, 12(1), 75–79. Kobayashi, T. (2010). Bridging social capital in online communities: Heterogeneity and social tolerance of online game players in Japan. Human Communication Research, 36(4), 546–569. Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., & Crawford, A. (2002). Internet paradox revisited. Journal of Social Issues, 58(1), 49–74. Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., & Scherlis, W. (1998). Internet paradox: A social technology that reduces social involvement and psychological well-being? American Psychologist, 53(9), 1017–1031. Ledbetter, A. M., Mazer, J. P., DeGroot, J. M., Meyer, K. R., Yuping, M., & Swafford, B. (2011). Attitudes toward online social connection and self-disclosure as predictors of Facebook communication and relational closeness. Communication Research, 38(1), 27–53. Lee, S. J. (2009). Online communication and adolescent social ties: Who benefits more from internet use? Journal of Computer-Mediated Communication, 14(3), 509–531. Lee, W., & Kuo, E. C. (2002). Internet and displacement effects: Children’s media use and activities in Singapore. Journal of Computer-Mediated Communication, 7(2). . Leung, L. (2002). Loneliness, self-disclosure, and ICQ (‘‘i seek you’’) use. CyberPsychology & Behavior, 5(3), 241–251. Lin, C.-P. (2011). Assessing the mediating role of online social capital between social support and instant messaging usage. Electronic Commerce Research and Applications, 10(1), 105–114. Marlow, C. (2006). Linking without thinking: Weblogs readership and online social capital formation. In Proceedings of the International Communication Association 2006. Nie, N., & Erbring, L. (2002). Internet and society: A preliminary report. IT&SOCIETY, 1(1), 275–283. Nie, N., Hillygus, D., & Erbring, L. (2002). Internet use, interpersonal relations, and sociability: A time diary study. In B. Wellman & C. A. Haythornthwaite (Eds.), The Internet in everyday life (pp. 215–243). Malden, Mass.: Blackwell Publ.. Parks, M. R. (2007). Personal relationships and personal networks. Mahwah, N.J: Erlbaum. Preacher, K. J., & Hayes, A. F. (2008). Contemporary approaches to assessing mediation in communication research. In A. F. Hayes, M. D. Slater, & L. B. Snyder

189

(Eds.), The Sage sourcebook of advanced data analysis methods for communication research (pp. 13–54). Los Angeles, Calif.: SAGE Publ.. Putnam, R. D. (1995). Bowling alone: America’s declining social capital. Journal of Democracy, 6(1), 65–78. Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster. Regan, P. C. (2011). Close relationships. New York: Routledge. Sciandra, M. (2011). Social capital and social network sites: An empirical analysis of European high school students. Italian Journal of Sociology of Education, 3(3), 52–68. Sheldon, P. (2009). ‘‘I’ll poke you. You’ll poke me!’’ Self-disclosure, social attraction, predictability and trust as important predictors of Facebook relationships. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 3(1), article 1. Shen, C. (2010). The patterns, effects and evolution of player social networks in online gaming communities. . Siitonen, M. (2007). Social interaction in online multiplayer communities. Jyväskylä, Finnland: University Printing House. Skoric, M. M., & Kwan, G. C. E. (2011). Platforms for mediated sociability and online social capital: The role of Facebook and massively multiplayer online games. Asian Journal of Communication, 21(5), 467–484. Steinkuehler, C., & Williams, D. (2006). Where everybody knows your (screen) name: Online games as ‘‘third places’’. Journal of Computer-Mediated Communication, 11(4). . Stutzman, F., Vitak, J., Ellison, N. B., Gray, R., & Lampe, C. (2012). Privacy in interaction: Exploring disclosure and social capital in Facebook. In Proceedings of the sixth international AAAI conference on weblogs and social media. Taylor, T. L., & Witkowski, E. (2010). This is how we play it: What a Mega-LAN can teach us about games. In Proceedings of FDG (June 19–21, Monterey, CA, USA) (pp. 195–202). Trepte, S., Reinecke, L., & Juechems, K. (2012). The social side of gaming: How playing online computer games creates online and offline social support. Computers in Human Behavior, 28(3), 832–839. Valenzuela, S., Park, N., & Kee, K. F. (2009). Is there social capital in a social network site?: Facebook use and college students’ life satisfaction, trust, and participation. Journal of Computer-Mediated Communication, 14(2), 875–901. Wheeless, L. R. (1976). Self-disclosure and interpersonal solidarity: Measurement, validation, and relationships. Human Communication Research, 3(1), 47–61. Wheeless, L. R., & Grotz, J. (1976). Conceptualization and measurement of reported self-disclosure. Human Communication Research, 2(4), 338–346. Wheeless, L. R., & Grotz, J. (1977). The measurement of trust and its relationship to self-disclosure. Human Communication Research, 3(3), 250–257. Williams, D. (2006a). Groups and goblins: The social and civic impact of an online game. Journal of Broadcasting and Electronic Media, 50(4), 651–670. Williams, D. (2006b). On and off the ‘net: Scales for social capital in an online era. Journal of Computer-Mediated Communication, 11(2). . Williams, D. (2007). The impact of time online: Social capital and cyberbalkanization. CyberPsychology & Behavior, 10(3), 398–406. Williams, D., Ducheneaut, N., Xiong, L., Zhang, Y., Yee, N., & Nickell, E. (2006). From tree house to barracks: The social life of guilds in World of Warcraft. Games and Culture, 1(4), 338–361. Williams, D., Yee, N., & Caplan, S. (2008). Who plays, how much, and why? Debunking the stereotypical gamer profile. Journal of Computer-Mediated Communication, 13(4), 993–1018. Yee, N. (2006a). Motivations for play in online games. CyberPsychology & Behavior, 9(6), 772–775. Yee, N. (2006b). The demographics, motivations and derived experiences of users of massively-multiuser online graphical environments. Presence: Teleoperators and Virtual, 15(3), 309–329.